On 06/08/15 01:42, Shahin Ansari wrote: > Thank you so much for the email. I tried all of the switches you suggested, > and I wanted to share the results with you in hope that you would validate > what I think is the issue. Also if you can suggest possible remedies, I > would greatly appreciate it. I started with... > > Initial state: > LOCAL_CFLAGS := -march=armv7-a -mtune=cortex-a15 -D__ASSEMBLY__ > -mabi=aapcs-linux -gdwarf-2 -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs > -fno-strict-aliasing -fno-common -Werror-implicit-function-declaration > -Wno-format-security -fno-delete-null-pointer-checks -Os > -Wno-maybe-uninitialized -marm -fno-dwarf2-cfi-asm -fstack-protector > -mabi=aapcs-linux -mno-thumb-interwork -funwind-tables > -D__LINUX_ARM_ARCH__=7 -mcpu=cortex-a15 -msoft-float -Uarm > -Wframe-larger-than=1024 -Wno-unused-but-set-variable -fomit-frame-pointer > -g -Wdeclaration-after-statement -Wno-pointer-sign -fno-strict-overflow > -fconserve-stack -DCC_HAVE_ASM_GOTO > > try 1 > I tried removing -march statement out and rely on -mcpu option. So the flags > I used were: > > LOCAL_CFLAGS := -mtune=cortex-a15 -D__ASSEMBLY__ -mabi=aapcs-linux -gdwarf-2 > -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing > -fno-common -Werror-implicit-function-declaration -Wno-format-security > -fno-delete-null-pointer-checks -Os -Wno-maybe-uninitialized -marm > -fno-dwarf2-cfi-asm -fstack-protector -mabi=aapcs-linux -mno-thumb-interwork > -funwind-tables -D__LINUX_ARM_ARCH__=7 -mcpu=cortex-a15 -msoft-float -Uarm > -Wframe-larger-than=1024 -Wno-unused-but-set-variable -fomit-frame-pointer > -g -Wdeclaration-after-statement -Wno-pointer-sign -fno-strict-overflow > -fconserve-stack -DCC_HAVE_ASM_GOTO > > result: > [armeabi-v7a] AsmFilter : hello_tz <= main_normal.s > [armeabi-v7a] Assembly : hello_tz <= main_normal.filtered.s > ./obj/local/armeabi-v7a/objs-debug/hello_tz/main_normal.filtered.s: > Assembler messages: > ./obj/local/armeabi-v7a/objs-debug/hello_tz/main_normal.filtered.s:16: > Error: selected processor does not support ARM mode `smc #0' > make: *** [obj/local/armeabi-v7a/objs-debug/hello_tz/main_normal.o] Error 1 > > try 2 > I tried to use -march=armv7-a+sec and removed the -mcpu=cortex-a15 statment > as follows: > > LOCAL_CFLAGS := -march=armv7-a+sec -mtune=cortex-a15 -D__ASSEMBLY__ > -mabi=aapcs-linux -gdwarf-2 -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs > -fno-strict-aliasing -fno-common -Werror-implicit-function-declaration > -Wno-format-security -fno-delete-null-pointer-checks -Os > -Wno-maybe-uninitialized -marm -fno-dwarf2-cfi-asm -fstack-protector > -mabi=aapcs-linux -mno-thumb-interwork -funwind-tables > -D__LINUX_ARM_ARCH__=7 -msoft-float -Uarm -Wframe-larger-than=1024 > -Wno-unused-but-set-variable -fomit-frame-pointer -g > -Wdeclaration-after-statement -Wno-pointer-sign -fno-strict-overflow > -fconserve-stack -DCC_HAVE_ASM_GOTO > > Result: > sansari@ubuntu:~/android/gas$ ndk-build > jni/Android.mk:6: This is in NDK_ROOT /home/sansari/android/android-ndk-r10d > [armeabi-v7a] Compile arm : hello_tz <= main_normal.c > arm-linux-androideabi-gcc: error: unrecognized argument in option > '-march=armv7-a+sec' > arm-linux-androideabi-gcc: note: valid arguments to '-march=' are: armv2 > armv2a armv3 armv3m armv4 armv4t armv5 armv5e armv5t armv5te armv6 armv6-m > armv6j armv6k armv6s-m armv6t2 armv6z armv6zk armv7 armv7-a armv7-m armv7-r > armv7e-m armv8-a iwmmxt iwmmxt2 native > make: *** [obj/local/armeabi-v7a/objs-debug/hello_tz/main_normal.s] Error 1 > > Try 3 > I changed armv7-a+sec to armv7ve as follows: > > LOCAL_CFLAGS := -march=armv7ve -mtune=cortex-a15 -D__ASSEMBLY__ > -mabi=aapcs-linux -gdwarf-2 -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs > -fno-strict-aliasing -fno-common -Werror-implicit-function-declaration > -Wno-format-security -fno-delete-null-pointer-checks -Os > -Wno-maybe-uninitialized -marm -fno-dwarf2-cfi-asm -fstack-protector > -mabi=aapcs-linux -mno-thumb-interwork -funwind-tables > -D__LINUX_ARM_ARCH__=7 -msoft-float -Uarm -Wframe-larger-than=1024 > -Wno-unused-but-set-variable -fomit-frame-pointer -g > -Wdeclaration-after-statement -Wno-pointer-sign -fno-strict-overflow > -fconserve-stack -DCC_HAVE_ASM_GOTO > > Result: > > jni/Android.mk:6: This is in NDK_ROOT /home/sansari/android/android-ndk-r10d > [armeabi-v7a] Compile arm : hello_tz <= main_normal.c > arm-linux-androideabi-gcc: error: unrecognized argument in option > '-march=armv7ve' > arm-linux-androideabi-gcc: note: valid arguments to '-march=' are: armv2 > armv2a armv3 armv3m armv4 armv4t armv5 armv5e armv5t armv5te armv6 armv6-m > armv6j armv6k armv6s-m armv6t2 armv6z armv6zk armv7 armv7-a armv7-m armv7-r > armv7e-m armv8-a iwmmxt iwmmxt2 native > make: *** [obj/local/armeabi-v7a/objs-debug/hello_tz/main_normal.s] Error 1 > > I am starting to agree that the toolchain I am using may not support > security extension. Do you think I should look into building my own > toolchain? I have little experience with this, but if it fixes my issue I > would do it. Do you think the reason for "unrecognized argument errors above > is lack of support by the toolchain? I think the toolchain I use is provided > my ndk. It seems to me NDK simply passes the files and switches to gcc; I do > not see what sort of support may be required here. Do you know of other > toolchains I can use? I know codesourcery makes toolchains? Do you know of a > toolchain that supports security extensions, and provides a choice between > arm and gnu syntax perhaps? > > Could this have anything to do with the version of gcc I run? Here is what I > have: > sansari@ubuntu:~/android/gas$ gcc -v > Using built-in specs. > COLLECT_GCC=gcc > COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper > Target: x86_64-linux-gnu > Configured with: ../src/configure -v --with-pkgversion='Ubuntu > 4.9.2-10ubuntu13' --with-bugurl=file:///usr/share/doc/gcc-4.9/README.Bugs > --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr > --program-suffix=-4.9 --enable-shared --enable-linker-build-id > --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix > --with-gxx-include-dir=/usr/include/c++/4.9 --libdir=/usr/lib --enable-nls > --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug > --enable-libstdcxx-time=yes --enable-gnu-unique-object > --disable-vtable-verify --enable-plugin --with-system-zlib > --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo > --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64/jre > --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64 > --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.9-amd64 > --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar > --enable-objc-gc --enable-multiarch --disable-werror --with-arch-32=i686 > --with-abi=m64 --with-multilib-list=m32,m64,mx32 --enable-multilib > --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu > --host=x86_64-linux-gnu --target=x86_64-linux-gnu > Thread model: posix > gcc version 4.9.2 (Ubuntu 4.9.2-10ubuntu13) This is your host compiler, not the cross compiler for android. You'll need to run arm-linux-androideabi-gcc, which may be buried somewhere deep within the android build system. R. > > Please let me know if I should perhaps upgrade it. > > -----Original Message----- > From: gcc-help-owner@xxxxxxxxxxx [mailto:gcc-help-owner@xxxxxxxxxxx] On > Behalf Of Richard Earnshaw > Sent: Friday, July 31, 2015 10:01 AM > To: Shahin Ansari <shansari@xxxxxxxxxxx>; gcc-help@xxxxxxxxxxx > Subject: Re: what options to pass to gcc for support of arm security > extensions > > My apologies, my statement this morning was not quite accurate. > > SMC is only implemented as part of the security extensions feature. As such > it is not part of the base-line ARMv7-A architecture. > > To enable this feature you need to specify -march=armv7-a+sec (provided your > toolchain supports that). You should also be able to enable it if you build > with -march=armv7ve (which includes SEC by default), or if you take the > -march= statement out of the list of options below and rely on picking up > the architecture from the -mcpu statement (your current command line options > are slightly in conflict, since Cortex-A15 implements the ARMv7VE version of > the architecture). > > Note that SMC will be treated as undefined if executed from user mode; user > code cannot call directly into the monitor. But that's a different issue. > > Once again, apologies for misleading you earlier. > > R. > > On 31/07/15 14:10, Shahin Ansari wrote: >> Thanks. The following are the flags I am using; I do not see what I am >> doing wrong. Would you have a look? >> LOCAL_CFLAGS := -march=armv7-a -mtune=cortex-a15 -D__ASSEMBLY__ >> -mabi=aapcs-linux -gdwarf-2 -Wall -Wundef -Wstrict-proto types >> -Wno-trigraphs -fno-strict-aliasing -fno-common >> -Werror-implicit-function-declaration -Wno-format-security -fno-de >> lete-null-pointer-checks -Os -Wno-maybe-uninitialized -marm >> -fno-dwarf2-cfi-asm -fstack-protector -mabi=aapcs-linux -mno >> -thumb-interwork -funwind-tables -D__LINUX_ARM_ARCH__=7 >> -mcpu=cortex-a15 -msoft-float -Uarm -Wframe-larger-than=1024 -Wn >> o-unused-but-set-variable -fomit-frame-pointer -g >> -Wdeclaration-after-statement -Wno-pointer-sign -fno-strict-overflow - >> fconserve-stack -DCC_HAVE_ASM_GOTO >> >> -----Original Message----- >> From: gcc-help-owner@xxxxxxxxxxx [mailto:gcc-help-owner@xxxxxxxxxxx] >> On Behalf Of Richard Earnshaw >> Sent: Friday, July 31, 2015 5:46 AM >> To: Shahin Ansari <shansari@xxxxxxxxxxx>; gcc-help@xxxxxxxxxxx >> Subject: Re: what options to pass to gcc for support of arm security >> extensions >> >> On 30/07/15 20:29, Shahin Ansari wrote: >>> Greetings- >>> According to the ARM documentation the SMC call should be supported >>> at >>> ARMv6 and above. However, I get the error message below when I try to >>> use >> it: >>> >> >> Almost, it's ARMv6Z that introduced the SMC instruction. It will >> certainly be recognized for ARMv7-a and above. >> >> GCC uses the -mcpu= (or if that's not specified the -march=) options >> from the command line to tell the assembler what instructions should >> be permitted. >> >> R. >> >
